Dynamic DICOM Endpoints for Diomede

Teleradiology in underserved regions such as rural Alaska depends on reliable, low-latency transfer of large DICOM medical imaging files between remote clinics and reading centers. Today, these transfers are governed by static configurations - a fixed IP address, port, and Application Entity (AE) Title standardized decades ago. This breaks silently when SCP nodes restart, IPs change, or storage queues fill. This proposal introduces Dynamic DICOM Endpoints, an intelligent orchestration layer integrated with the open-source Orthanc PACS server and Docker containers on the Cloud. A FastAPI Orchestrator backed by a real-time Redis registry continuously monitors the health, latency, queue depth, and storage capacity of every Orthanc node in a multi-region deployment, selecting the optimal routing destination in milliseconds and replacing fragile static pipes with a self-healing, load-aware mesh. Deliverables include a four-region Docker Compose simulation environment; a Python telemetry daemon; a pluggable scoring algorithm; a thin Edge Agent bridging legacy scanners (GE, Siemens); a comprehensive test and benchmarking suite validating routing correctness, failover behavior, and transfer performance; and a documentation suite. All code will be contributed as open-source to the Diomede repository, directly improving imaging infrastructure for remote and resource-constrained healthcare settings. As part of my pre-proposal engagement with the host organization, I submitted Issues #59, #79, and #85, along with the corresponding PRs #60, #80, and #86, to the Diomede repository. This contribution demonstrates my familiarity with the project's codebase and development workflow, and establishes a collaborative relationship with the maintainers ahead of the GSoC coding period.